Outi Vesakoski (University of Turku) & Michael Dunn (Uppsala University):

The role of historical linguistics in the interdisciplinary study of the human past

Language carries information about past events, and a better understanding of historical linguistics and results therein would allow better usage of such data in interdisciplinary studies of the human past. 

A key source of evidence about language change comes from  sound change alternations, i.e. regular ‘mutations’ in the phonetic systems of a language, so that all words containing a particular sound in a particular context change more-or-less simultaneously in the same way.  As sound changes themselves change the contexts that other sound changes appear in it is possible to determine the order of changes, and thus track ancestral stages of the language backwards through history.

Sound changes are not transparent, but a linguist can tell that “hal” in Hungarian and “kol” in Mari both go back to proto-Uralic “kala” ‘fish’. Such words descended from the same ancestral form are called cognates, which are equivalent to homologies in biological data. The raw materials for phylogenetic linguistics are cognate assessments, made of most conserved part of the lexicon. Most quantitative phylogenetic methods take a Bayesian approach to building family trees, which can  be used statistically for e.g. evaluating different hypotheses proposed for the history of the language family. Bayesian approaches also provide the possibility of timing branching events in the family. Probabilistic timings are useful in interdisciplinary analyses for many reasons: they allow integration of information from many different sources, and they provide material for modelling uncertainty. Besides the timing of the Uralic tree, we also present the methods underlying chronological analyses — a review of literature explicating the absolute timing of the Uralic branching events (Maurits et al. in press).

Part of the lexicon is inherited, but part is borrowed horizontally from languages in contact. Linguistic contacts can be an important data source for historical inference because they can function as proxies of population contacts. Furthermore, loan word layers can give relative timing —  e.g. Indo-Iranian loans entered the Uralic languages earlier than the Germanic loans. However, etymological studies are laborious. To overcome this, we tested if a study of etymologiocally well understood ‘basic vocabulary’ (a standard list of 100 - 200 meanings) would reveal the essential loan word in representative manner. This gives an effective proxy measure for  prehistoric linguistic contacts on lesser studied languages (de Heer et al. manuscript). 

In this introduction we aim to introduce the linguistic comparative method to a non-linguistic audience and on the other hand, to introduce computational phylogenetic approach to non-practitioners. For the interdisciplinary audience we will present several case studies showing how historical linguistics can contribute to research on holistic human past.